|Differences in HIV spread in four sub-Sharan African cities (UNAIDS, 1999, 9 p.)|
Lusaka, 14 September 1999
Summary of the multi-site study*
* The full report of the multi-site study has been submitted for publication.
In many large towns in Central, East and southern Africa the HIV prevalence rate among pregnant women currently exceeds 30%, while in the cities and towns of most of West Africa fewer than 10% of pregnant women are infected. What accounts for these enormous variations? Some urban populations admittedly have a longer-standing epidemic than others, but the time factor alone does not explain all the differences. Clearly, HIV has been spreading at different rates in different populations.
To understand more about this differential rate of spread, a multi-site study was carried out in four African towns with differing HIV prevalence trends in pregnant women. Kisumu (Kenya) and Ndola (Zambia), in Central/East Africa, were selected as the towns with high HIV prevalence, while the low-prevalence towns in West Africa were Cotonou (Benin) and YaoundCameroon).
Between June 1997 and March 1998, surveys were carried out in all four sites to compare:
· sexual behaviour patterns - e.g. age of sexual initiation and marriage, number of sex partners, contacts with prostitutes - which may determine the probability of exposure to an infected partner
· co-factors - e.g. condom use, other sexually transmitted diseases, male circumcision - which may determine the probability of HIV transmission during intercourse.
Ethical approval for the study was obtained from the national ethics committee in each of the countries where the study took place, as well as from other ethics committees. The general population survey in each town aimed at a random sample size of 1000 men and 1000 women aged 15 to 49 years. Households were visited by a team of interviewers and nurses or doctors. The survey of sex workers was preceded by mapping all places where prostitutes could be found and recording the number of sex workers present in each location at the time of the study teams visit, so that a representative sample of about 300 prostitutes could be selected in each town.
After giving their informed oral consent, study participants were interviewed on their socio-demographic characteristics and sexual behaviour, using a standardized questionnaire, and were asked to give a blood sample and urine sample. HIV testing was done anonymously, but the result was linked to the interview data and to the results of the laboratory tests for HIV and for the presence of sexually transmitted diseases (STDs) including syphilis, gonorrhoea, genital herpes (HSV-2) and trichomonas infection. Study participants who wished to know their HIV serostatus were referred for pre- and post-test counselling and re-testing, free of charge. Study participants with symptoms and/or signs suggestive of an STD were treated.
Prevalence of HIV infection
The prevalence of HIV infection in men (aged 15-49) was 3% in Cotonou, 4% in Yaound20% in Kisumu and 23% in Ndola. For women, the respective prevalence rates were 3%, 8%, 30% and 32%.
In all sites, except Cotonou, overall prevalence was significantly higher in women than in men. The largest male/female divergence was found among 15-19-year-olds in Kisumu and Ndola. In Kisumu, over 3% of male teenagers were infected, compared with 23% of female teenagers. In Ndola, HIV prevalence in this age group was 4% in boys and 15% in girls.
The prevalence of HIV infection among sex workers was extremely high in the two high-prevalence sites (Kisumu, almost 75%; Ndola, 68%) compared with Yaound33%), although the prevalence in Cotonou (over 57%) was also quite high. Across these four sites 40% of the prostitutes were 24 years old or younger, with a figure of 56% in Ndola. Almost a quarter of prostitutes in Ndola were under 20.
In the two low-prevalence sites and in Kisumu, more than 70% of circulating HIV-1 strains belonged to subtype A. In Ndola practically all circulating strains were subtype C. The biggest range of strains was see in Yaoundwhere subtypes A, D, E, F, G and H were found, as well as strains belonging to group O.
Sexual behaviour characteristics
Sexual debut and marriage
In all sites, between 97% and 99% of men and of women 20 years old and above reported having had sexual intercourse. However, among 15-19-year-olds the proportion of those who were sexually active ranged from under 50% in Cotonou to just over 70% in Kisumu. For both males and females, age at first sexual intercourse was lowest in Kisumu and highest in Cotonou. Significantly more girls in Kisumu and Ndola had their sexual debut before age 15 than in the two low-prevalence sites.
The percentage of men and women who were currently-married or ever-married was significantly higher in Kisumu and Ndola than in the low-prevalence sites. Similarly, age at first marriage was significantly lower for both men and women in Kisumu and Ndola than in Cotonou and Yaound
Rate of partner change
The highest rate of partner change was found in Yaoundwhere men reported an average of 10 lifetime partners (compared with 5 in Kisumu and 4 in Cotonou and Ndola) and women reported 3 (compared with 2 in the other three sites). Never-married men and women in Yaoundeported the largest numbers of non-spousal partners in the year preceding the interview, followed by men and women in Kisumu. Currently-married men in the high-prevalence sites reported fewer non-spousal (extramarital) partners than married men in Cotonou and YaoundExtramarital relationships were reported by very few currently-married women (1-3%) except in Yaoundwhere 12% of women reported at least one such relationship in the year before the interview.
Age mixing between non-spousal partners
Age differences between non-spousal partners were similar in Cotonou, Yaoundnd Ndola, but significantly smaller in the high-prevalence site, Kisumu. The age difference between non-spousal partners, as reported by men, was 4 in Cotonou and Yaound3 in Kisumu and 4 in Ndola, the male partner always being older. However, among married - teenage girls living in Kisumu and Ndola, the older their husband (and the larger the age difference between the partners), the higher their risk of being HIV-infected.
Male contacts with sex workers
The proportion of men reporting at least one contact with a sex worker in the past year was almost 7% for Cotonou, 13% for Yaoundalmost 7% for Kisumu and 11% for Ndola. When analysis was restricted to men who reported having non-spousal partnerships in the year before interview, the percentage having contact with a prostitute was significantly higher in Ndola (over 32%) than in Cotonou and in Kisumu (around 15%) and Yaoundalmost 20%).
While HIV prevalence among sex workers tended to be higher in the high-prevalence sites, as mentioned earlier, the HIV rate in Cotonou sex workers was also quite high. However, use of a condom with the most recent client was reported more often by sex workers in Cotonou than in any other site. Of the HIV-infected sex workers in Cotonou 64% reported condom use with their last client, compared with 349% in Yaound50% in Kisumu and 30% in Ndola. When the data on HIV prevalence and condom use with clients are examined together, it is clear that sex work clients in Ndola and Kisumu are more exposed to HIV infection than clients in the low-prevalence towns.
A relatively low proportion of men in all sites - ranging from 21% in Cotonou to 25% in Ndola - reported that they often or always used condoms with non-spousal partners. Women in the low-prevalence towns reported less condom use than women in the high-prevalence sites.
Co-factors in HIV transmission
Syphilis, an STD that causes genital ulcers, was by far most prevalent in Ndola, followed by YaoundSerologic evidence of current or previous genital herpes (HSV-2 infection) - another ulcer-producing STD - was more common in the high-prevalence sites than in the other two sites. Taken together, the prevalence of the two ulcerative STDs was higher in the towns with high HIV prevalence. The differences between the sites were especially marked among young people under 25. In all four sites HSV-2 infection was strongly associated with an increased risk of HIV infection in both men and women.
The prevalence of gonorrhoea was generally low and variable. Chlamydial infection was more common in younger age groups, especially in girls, and the highest prevalence was in Yaoundfollowed by Kisumu. The highest rates of trichomonas infection were found in women in Kisumu and Ndola (29% and 34%).
In Cotonou and Yaoundearly all men reported being circumcised, while in Ndola only 10% of men were circumcised. In Kisumu, where the overall percentage of circumcised men was less than 30%, HIV prevalence was below 8% in men circumcised before their sexual debut and 25% in uncircumcised men. In multivariate analysis, being circumcised was associated with a lower risk of HIV infection.
In our study the only differences in sexual behaviour that distinguished Kisumu and Ndola from Cotonou and Yaoundoncerned the age at which girls became sexually active, and the age at which men and women first married. Compared with the low-prevalence sites, in Kisumu and Ndola females were younger at sexual debut and both sexes married earlier. In Kisumu, moreover, teenage girls whose husbands were older were more likely to be HIV-infected. High rates of partner change, contacts with sex workers, and concurrent sexual partnerships, were not reported more systematically in the high-prevalence than in the low-prevalence sites.
From these data, it would be difficult to argue that the divergence in the rate of HIV spread between the East African and West African sites can be explained solely by differences in sexual behaviour.
Nor was there evidence that differences in circulating strains of HIV-1 are a major factor in the rate of spread of HIV. Subtype A was the most prevalent subtype in both of the low-prevalence sites as well as in one high-prevalence site.
However, the study did find associations between higher HIV rates and two biological co-factors, suggesting that these might have contributed to a higher probability of HIV transmission during sexual intercourse in the high-prevalence sites.
Previous studies had already established that HIV transmission is enhanced in the presence of another STD, particularly an STD causing genital ulcers. The multi-site study found that significantly more people in Kisumu and Ndola, especially in the 15-24 age bracket, had serologic evidence of current or previous infection with an ulcerative STD (syphilis and genital herpes).
Secondly, the study found large differences between the sites in the percentage of men circumcised. In Cotonou and Yaoundcities where over 97% of men are circumcised, lower HIV rates were found than in Kisumu and Ndola, where the percentage of men circumcised ranges from 10% to less than 30%. In the high HIV prevalence sites, moreover, close to 16% of young men under 25 had an ulcerative STD, whereas in the low-prevalence sites the figure was 6%-8%. Lack of circumcision was thus associated with higher rates of both HIV and ulcerative STDs, the latter in turn possibly raising the HIV transmission risk.
Finally, how can one explain the dramatically high prevalence of HIV infection (15%-23%) in girls under 20 in Kisumu and Ndola? Unmarried girls in these cities reported a median of 1 to 1.5 lifetime sex partners, an estimated 10-12% of whom were HIV infected. A separate qualititative study carried out in Kisumu has found evidence of early sex with older partners (men aged 25 and above) among teenage girls, both married and unmarried. This suggests that the teenagers in the multi-site study may have under-reported their number of partners, especially non-spousal partners older than 25. Partners like these are likely to have exposed the girls to the virus, since HIV prevalence in Kisumu and Ndola among men aged 25 and over ranged from 26% to 40%.
Even when these factors are taken into account, however, it is hard to explain the high HIV prevalence in female teenagers. For the girls to have become infected so soon after their sexual debut as a result of relatively few exposures to an infected partner, HIV transmission co-factors must be part of the explanation. In this connection it is important to recall that almost 50% of the 15-19-year-old girls in Kisumu and Ndola had been exposed to the virus that causes genital herpes and almost 16% had syphilis.
In conclusion, differences in the rate of HIV spread between the East African and West African cities studied cannot be explained away by differences in sexual behaviour alone. In fact, behavioural differences seem to be outweighed by differences in HIV transmission probability.
The implications of our findings for the prevention measures are complex. For instance, if further studies prove that male circumcision is really protective, it would be important to assess carefully the benefits as well as the practical risks of the procedure under field conditions, including the risks of infection and haemorrhage. With respect to the ulcerative STDs, while syphilis is curable there are no drugs that can cure or reduce the transmission of genital herpes - a lifelong viral infection. Herpes can, however, be prevented through condom use, and research is in progress to develop a vaccine.
On the other hand, our findings should not be interpreted as a denial of the important role that sexual behaviour change and improved STD care can play in curbing the HIV epidemic. First of all, even the low HIV prevalence sites studied have considerable rates of HIV infection, and even in the high-prevalence sites barely a quarter of men reported that they often or always used condoms with non-spousal partners. The risk behaviours identified in all four sites call for scaled-up and sustained action in the areas of condom promotion, education for safer sexual behaviour, and diagnosis and treatment of the curable STDs.
Secondly, given the high levels of HIV infection in young women in Kisumu and Ndola, effective interventions are urgently needed to decrease their vulnerability. Girls must be made aware that they run an enormous risk of becoming infected with ulcerative STDs, HIV, or both, during their first few exposures to sex, especially with an older man, who is far more likely to be infected than boys their own age. Girls should learn the necessary life-skills to stand up to demands for early, unwanted or unsafe intercourse.
Above all, since men still play the dominant role in deciding whether and under what circumstances sex will take place, priority must be given to sexual behaviour change programmes aimed at them. Social pressure should be put on older men to avoid forcing or coercing young girls into sex, or enticing them with sugar-daddy gifts. Cross-generational sex exposes girls to lethal risk and helps drive the HIV epidemic.
Members of the Study Group on Heterogeneity of HIV Epidemics in African Cities are: A Buvcoordinator), M Laga, E Van Dyck, W Janssens, L Heyndricks (Institute of Tropical Medicine, Belgium); S Anagonou (Programme national de Lutte contre le SIDA, Benin); M Laourou (Institut national de Statistiques et dAnalyses nomiques, Benin); L Kanhonou (Centre de Recherche en Reproduction humaine et en Dgraphie, Benin); Evina Akam, M de Loenzien (Institut de Formation et de Recherche dgraphiques, Cameroon); S-C Abega (Universitatholique dAfrique Centrale, Cameroon); Zekeng (Programme de Lutte contre le SIDA, Cameroon); J Chege (The Population Council, Kenya); V Kimani, J Olenja (University of Nairobi, Kenya); M Kahindo (National AIDS/STD Control Programme, Kenya); F Kaona, R Musonda, T Sukwa (Tropical Diseases Research Centre, Zambia); N Rutenberg (The Population Council, USA); B Auvert, E Lagarde (INSERM U88, France); B Ferry, N LydiCentre frans sur la population et le dloppement, France); R Hayes, L Morison, H Weiss, J Glynn (London School of Hygiene & Tropical Medicine, UK); NJ Robinson (Glaxo Wellcome, UK); (M Cara(UNAIDS, Switzerland).
For more information, please contact Anne Winter, UNAIDS, (+41 22 791.4577)